@phdthesis{69cf630a-c4f4-4541-9e85-a37dceef25ae,
  abstract     = {{In this thesis machine learning applications to modern particle accelerators are developed and tested at a number of facilities. The focus is on improving and advancing longitudinal diagnostics, with a major project focusing on non-destructive predictions of beam distributions in longitudinal phase space. Applications are developed using data collected from MAX IV, FERMI, SwissFEL and FLASH. Beyond this, the use of machine learning for optimization of linear accelerators is implemented with focus on complex tasks such as longitudinal phase space and emittance optimization. <br/><br/>Further,  we advance the use of longitudinal phase space diagnostics in connection with other studies focused on longitudinal beam dynamics. Projects include investigations into the high-order shaping of the longitudinal phase space and the further development of beam compression schemes, including experimental measurements of arc-like bunch compressors and investigations of their inherent advantages.}},
  author       = {{Lundquist, Johan}},
  isbn         = {{978-91-6858-031-9}},
  keywords     = {{Accelerators; Beam diagnostics; Machine Learning; Beam dynamics; Artificial neural network; Free electron laser; supervised learning; transverse deflecting structure}},
  language     = {{eng}},
  month        = {{09}},
  publisher    = {{Lund University}},
  school       = {{Lund University}},
  title        = {{Machine Learning for Longitudinal Monitoring and Control of Ultra-Bright Electron Beams}},
  url          = {{https://lup.lub.lu.se/search/files/261073963/Thesis_JohanL_FinalPrint_Kappa.pdf}},
  year         = {{2026}},
}